How do I compare two websites objectively?
Benchmarking against a competitor without fooling yourself — what is fair to compare, and what is not comparable at all.
Measure both with the same instrument at roughly the same time, and compare findings rather than headline scores. Configuration — headers, transport, crawler access, accessibility, metadata — compares cleanly. Content quality, authority and traffic do not, and any tool implying otherwise is guessing.
What compares cleanly
These are objective properties of what a server sends. Two sites can be placed side by side without qualification.
- Security headers. Present or absent, with values. No interpretation required.
- Transport. Protocol version, cipher, certificate validity and expiry.
- Crawler access. Which bots each site allows, resolved from the actual rules.
- Indexability. Canonical tags, noindex directives, sitemap presence.
- Accessibility basics. Alt attributes, labels, heading order, contrast — all mechanically checkable.
- Metadata hygiene. Titles, descriptions, structured data validity.
- Third-party footprint. How many external origins each page pulls in.
What does not compare
- Content quality. A specialist page serving fifty people perfectly will look thin next to a generalist page serving nobody in particular. Word counts and readability scores do not capture fitness for purpose.
- Authority. Not measurable from the site itself at all. Anything claiming to score it is modelling, and the model is proprietary and unfalsifiable.
- Traffic and rankings. Not observable from a site. Estimates are estimates.
- Whether the business is doing well. Worth saying plainly, because a comparison table invites the inference.
Doing it properly
- Measure both sites with the same tool, close together in time. Configuration changes; a comparison across a fortnight is not a comparison.
- Compare the same URL type — homepage against homepage, product page against product page. A homepage and a blog post are different objects.
- Read what was not covered for both. If one site blocked half the audit, the comparison is between different amounts of information.
- Compare findings, not scores. "They have HSTS and we do not" is actionable; "they score 84 and we score 79" is not.
- Check the undetermined lists on both sides before drawing any conclusion.
Step five catches the most common misreading: a competitor behind aggressive bot protection produces fewer findings, which is easy to mistake for a better-configured site.
Comparing several at once
for site in example.com competitor-a.com competitor-b.com; do
printf '%-24s %s\n' "$site" \
"$(curl -sf "https://outrings.com/api/v1/score?url=$site" | jq -r '.score')"
sleep 3
doneFor a structured comparison including the per-category gaps, the compare_websites MCP tool takes two to four sites and returns them ranked with the differences named.
The useful output
The point of a comparison is not a ranking. It is a list of things a peer does that you do not, which is a to-do list assembled by someone else's engineering effort — genuinely one of the cheapest sources of good ideas available. Read it that way and the exercise pays for itself even when you come out ahead.
What our audit reports about this
Every item below is measured directly, not inferred. Run it against your own site and the result names the exact rule or header responsible.
- Identical checks applied to every site, so a comparison is like for like by construction.
- A compare tool taking two to four sites and returning them ranked with per-category gaps named.
- Coverage and undetermined lists per site, so a partly blocked audit is not mistaken for a clean one.
- Deterministic scoring, so measuring the same two sites again produces the same relative result.
For agents and scripts, the same measurement is at
/api/v1/summary?url=yoursite.com —
see the API documentation.
Related questions
Is it legal to audit a competitor?
It sends ordinary HTTP requests and reads publicly served responses, which is what every browser and crawler does. It is not a vulnerability scanner, does not probe for weaknesses and does not attempt authentication.
Why does my competitor score higher with worse content?
Because the score measures configuration, not content quality. A well-configured empty template outscores an excellent site with missing headers. Read the findings rather than the number.
Should I copy everything they do?
Copy the specific configuration wins — headers, transport settings, metadata hygiene. Do not copy content strategy from a score, which says nothing about whether their content works.
How many sites should I compare against?
Three to five peers is plenty. Beyond that the findings repeat, and the exercise turns into collecting numbers rather than collecting things to fix.
Read next
How do I audit hundreds of websites at once?
Running a portfolio, a client list or a research sample through a measurement API without abusing anyone — including the sites you are measuring.
ReadWhat does a website score actually mean?
How a score out of 100 is built, what it can honestly tell you, and the four things it cannot.
ReadHow do I audit a website I don't own?
What is legitimate to measure on someone else's site, where the line is, and how to do it without being a nuisance.
ReadWhy do two SEO tools give my site different scores?
Different checks, different weights, different definitions of failure — and why comparing scores across tools is meaningless.
Read